Generative AI-Driven Semantic Communication Networks: Architecture, Technologies, and Applications

• DOI: 10.1109/tccn.2024.3435524
• Publication Date: 2024-07-29T19:46:16Z
• Authors: Chengsi Liang; Hongyang Du; Yao Sun; Dusit Niyato; Jiawen Kang; Dezong Zhao; Muhammad Ali Imran
• Keywords: 05 social sciences; Electrical Engineering and Systems Science - Signal Processing; 0506 political science

Generative artificial intelligence (GAI) has emerged as a rapidly burgeoning field demonstrating significant potential in creating diverse content intelligently and automatically.To support such intelligence-generated (AIGC) services, future communication systems must fulfill stringent requirements, including high data rates, throughput, low latency, while efficiently utilizing limited spectrum resources.Semantic (SemCom) been deemed revolutionary scheme to tackle this challenge by conveying the meaning of messages instead bit reproduction.GAI algorithms serve foundation for enabling intelligent efficient SemCom terms model pre-training fine-tuning, knowledge base construction, resource allocation.Conversely, can provide AIGC services with latency reliability due its ability perform semantic-aware encoding compression data, well knowledge-and context-based reasoning.In survey, we break new ground investigating architecture, wireless schemes, network management GAI-driven networks.We first introduce novel architecture networks, comprising plane, physical infrastructure, control plane.In turn, an in-depth analysis transceiver design semantic effectiveness calculation end-to-end systems.Subsequently, present innovative generation level strategies proposed update, sharing, ensuring accurate timely knowledge-based reasoning.Finally, explore several promising use cases, i.e., autonomous driving, smart cities, Metaverse, comprehensive understanding direction networks.